In the PCPT and REDUCE trials, finasteride and dutasteride significantly reduced the detection of prostate cancer (PCa) by 23 and 25% respectively, and thus established that 51-reductase inhibitors (5ARI) are the first class of drugs proven as chemopreventive agents for PCa. The actual efficacy of these drugs varies among individuals, even after considering compliance with the prescribed dosage. Better understanding of the basis for sensitivity to 5ARI chemoprevention will: 1) pave the way for development of better-tolerated and more effective agents that exploit this established mechanism, 2) allow clinicians to target responsive men, thus reducing the cost and morbidity of life-long dosage, and 3) validate specific tissue biomarkers as surrogate endpoints for Phase II trials. Our preliminary data based on direct DNA staining suggest that nuclear morphometric features that characterize 5ARI response may also define a field effect that predicts PCa in untreated men with negative biopsies. REDUCE, which was completed in 2009, provides a unique opportunity to address these questions because intermediate, on-study biopsy samples (mandatory at Years 2 and 4) were collected. Previous research indicated that 5ARIs (at higher doses for short periods) produce changes in benign prostate that resemble incomplete atrophy. Our overall goal is to apply cutting- edge imaging approaches, incorporating machine-learning for pattern recognition and multispectral analysis, to the development and validation of intermediate endpoint biomarkers in benign tissue that characterize the response to 5ARI chemoprevention as well as the risk of PCa among men with negative biopsies. We will obtain Year 2 slides and tissue blocks from a random sample of REDUCE participants with various outcomes. Aim 1: To determine the effects of dutasteride (vs. placebo) on both nuclear and architectural features in benign tissue. Aim 2: To determine whether a multivariable treatment-response score differs between subjects who develop PCa while on dutasteride and those who do not, and, Aim 3: To determine the magnitude of association between nuclear phenotype in benign biopsies, and subsequent risk of PCa in untreated men at elevated risk. Altogether, we will use 3 techniques to assess cytomorphology: a) a morphometric score based on nuclear size, shape and texture, b) expression -via quantum dot imaging - of p300 and nucleolin (two proteins with major effects on chromatin pattern and nuclear morphology) and c) mapping of architectural features (e.g., epithelial area and height) via trainable software. This will be the first work to relate the cytomorphological effects of a 5ARI, given at a chemopreventive dose level for a lengthy period, to subsequent cancer occurrence. The results will have implications for chemoprevention research and clinical practice, including the large number of U.S. men who remain at risk following a negative prostate biopsy.